Hello,

I am working with someone else's code applying a fmm.
The below code shows an example f output from the user written package (used previous to stata 15).
In stata 16, I can produce the same numbers (using the following:

fmm 2 ... regress var# ...

followed by and "estatlcmean" to get means and "estat lcprob " to get the pi1 and pi2.

However, I am unclear how to reproduce the "sigma1" and "sigma2" in the output table below. Is there a post estimation command I am missing here?

Thanks,


2 component Normal regression Number of obs = 11,102
Wald chi2(0) = .
Log pseudolikelihood = -90736.087 Prob > chi2 = .

(Std. Err. adjusted for 610 clusters in psu)
------------------------------------------------------------------------------
| Robust
var1| Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
component1 |
_cons | 2951.496 15.38723 191.81 0.000 2921.337 2981.654
-------------+----------------------------------------------------------------
component2 |
_cons | 2944.716 56.66418 51.97 0.000 2833.656 3055.776
-------------+----------------------------------------------------------------
/imlogitpi1 | 1.499674 .7441516 2.02 0.044 .041164 2.958185
/lnsigma1 | 6.213801 .0483252 128.58 0.000 6.119085 6.308516
/lnsigma2 | 6.682482 .1061851 62.93 0.000 6.474363 6.890601
------------------------------------------------------------------------------
sigma1 | 499.5965 24.14308 454.4487 549.2294
sigma2 | 798.2977 84.76731 648.3059 982.9916
pi1 | .8175259 .1110105 .5102896 .9506489
pi2 | .1824741 .1110105 .0493511 .4897104